KT

Kevin Tian

The University of Texas at Austin · United States

Unclaimed · AI-structured profileORCID OpenAlex

UT Austin theorist working on streaming PCA, differential privacy, and sampling algorithms for spin glasses

3

papers here

2

reviewed by Evolve

0

first-author

0

corresponding

—

works · OpenAlex

11

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output: insufficient dataOutputImpact: insufficient dataImpactFocus: 0/100FocusCollaboration: 46/100CollaborationActivity: 99/100ActivityReproducibility: 55/100Reproducibility

Output—

No output score is available yet: a normalized productivity measure over venues and career stage has not been computed for this profile. This axis would measure total, per-year, and venue-normalized publication output, beyond the raw yearly counts already on record.

Impact—

Impact cannot be scored yet: beyond an h-index of 11 and i10-index of 12, per-paper citations, venue tier, and downstream uptake are absent from the current record. This axis would measure citations, influence, and adoption of the work by other researchers.

Focus0/100

The focus score is 0, the floor of the scale: the record splits exactly evenly (0.5/0.5) between machine learning theory — streaming PCA and differential privacy — and theoretical computer science work on sampling, spin glasses, and statistical-computational gaps. This is breadth across two theory areas rather than scatter, since topic clusters such as stochastic gradient optimization (32) and graph complexity (24) stay within optimization and sampling methodology.

Collaboration46/100

Collaboration scores 46, mid-range: the record lists 9 distinct coauthors, a compact network centered on a UT Austin theory group. Recurring partners include Syamantak Kumar, Anming Gu, Chutong Yang, Purnamrita Sarkar, and Yusong Zhu, with Gautam Kamath also frequent. On the three Evolve-reviewed 2026 papers he appears in middle or last author position rather than as sole leader, consistent with small-team theory collaborations.

Activity99/100

Activity scores 99, near the ceiling: 29 works in the most recent three years (2024: 9, 2025: 8, 2026: 12 in the provided counts), with 2026 matching the historical single-year peak of 12 set in 2020. The record spans 2011 to 2026 with sustained output since 2017, indicating a currently intensive, preprint-driven phase centered on arXiv theory papers.

Reproducibility55/100

Reproducibility scores 55, the provided mean over 2 scored papers, both rated medium readiness by Evolve. The streaming PCA work rests on synthetic tests only, and the low-temperature sampling result near the Almeida-Thouless line is partly conditional — limitations that temper independent replication prospects. The third reviewed paper is not yet scored.

Research summary

Kevin Tian is a theoretical researcher at The University of Texas at Austin whose indexed work divides exactly evenly between machine learning theory — streaming PCA and differential privacy — and theoretical computer science, covering sampling algorithms, spin glasses, and statistical-computational gaps. The OpenAlex-derived record shows an h-index of 11, an i10-index of 12, and 29 works in the most recent three years, with an activity score of 99; annual counts run from single works in 2011 and 2013 up to peaks of 12 in 2020 and 2026. Topic clusters are led by stochastic gradient optimization (32), graph complexity and algorithms (24), and advanced optimization algorithms (15). Evolve has reviewed three 2026 arXiv preprints: a gap-free, general-rank analysis of Oja's algorithm with near-matching lower bounds, extending to differentially private PCA; sparsity-aware sampling frameworks giving any-temperature Sherrington-Kirkpatrick slice samplers and spike-and-slab guarantees; and a last-authored study of the provable-auditable privacy gap of DP-SGD. Collaboration is compact — 9 distinct coauthors including Syamantak Kumar, Anming Gu, and Purnamrita Sarkar — and reproducibility readiness is medium (mean 55 over 2 scored papers). The profile is name-resolved from public bibliographic data.

Collaboration pattern

Collaboration scores 46 on a compact network: 9 distinct coauthors, recurring partners including Syamantak Kumar, Anming Gu, Chutong Yang, Purnamrita Sarkar, and Yusong Zhu, plus Gautam Kamath. On all three Evolve-reviewed 2026 papers he takes a middle or last author position, suggesting contribution within a small UT Austin theory group rather than solo-led projects.

Research keywords

Streaming PCAOja's algorithmDifferential privacyRényi differential privacyIsing modelSherrington-Kirkpatrick modelMCMC mixing timeDown-up walkSpike-and-slab regressionStochastic gradient optimizationStatistical-computational gapsDP-SGD

Highlights

Timeline

  1. 2011Earliest year in the indexed record, with 1 work listed in the provided counts-by-year data.
  2. 2020Single-year output reaches 12 works, the highest annual count in the provided record.
  3. 2024Start of the recent high-activity stretch: 9 works this year, part of 29 works across 2024-2026.
  4. 2026Records 12 works, matching the 2020 peak, with three arXiv preprints reviewed by Evolve: gap-free streaming PCA, low-temperature Ising and sparse-regression sampling, and a DP-SGD privacy-gap study.

Papers in the Evolve catalog

Frequent co-authors

Is this you? Claim this profile

Enter the email you published with. We'll send a one-time claim link; after claiming you can edit the introduction or hide this page. Requests from a mismatched address go to manual review.

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

  • Bibliometrics are name-resolved from OpenAlex public data (identity: openalex, ORCID 0000-0003-0109-9201); Kevin Tian is a fairly common name, so the record may include or omit some works.
  • Output and impact dimensions are unscored: citation-level and venue-level aggregates beyond h-index 11 and i10-index 12 were not available.
  • The reproducibility mean (55) rests on only 2 scored papers; the third reviewed paper has no Evolve verdict, keywords, or reproducibility rating yet.
  • The frequent-coauthor list contained a duplicate entry (Yusong Zhu), treated once in this profile.
Kevin Tian — Researcher Profile | Evolve.Pub